Mundell-Fleming without the LM curve:
the exogenous interest rate in an open economy 
Franklin Serrano
IE/UFRJ
Ricardo Summa
IE/UFRJ
Abstract
In this paper we evaluate critically the popular Mundell-Fleming model from the
standpoint the exogenous interest rate heterodox approach. We criticize the
assumptions of exogenous money supply, "perfect" international capital markets and
inelastic exchange rate expectations. We show that in a more realistic framework none
of the main results of the Mundell-Fleming model on the relative effectiveness of fiscal
and monetary policies are valid, either in floating and fixed exchange rate regimes. We
conclude that ,within certain very asymmetric bounds, the central bank has the power
to determine the domestic interest rate exogenously even in open economy with free
capital mobility and that there is no automatic market mechanism to ensure the
automatic adjustment of the interest rate and exchange rate to sustainable levels.
Key words: Mundell-Fleming, Exogenous interest rate, Economic Policy.

The authors would like to thank for the very helpful comments of Prof. André Lourenço.
Mundell-Fleming without the LM curve: the exogenous interest rate in an open
economy
Franklin Serrano
Ricardo Summa
I.Introduction
Fortunately, there has been a growing number of orthodox and heterodox economists
that accept that the basic nominal interest rate is determined exogenously by the
central bank of a country and not by an adjustment between the "demand for money"
and an “exogenous” money supply.1. However, at the same time many economists still
think in terms of the Mundell-Fleming model (IS-LM-BP short-term external perfect
capital mobility) when discussing specific problems of open economy.2 As it is well
known, the Mundell-Fleming model is based on the idea that the interest rate is
determined by supply and demand for money and that money supply is exogenous, at
least in the first instance, because depending on the exchange rate regime it can be
made to vary endogenously. Moreover, in the Mundell-Fleming model the interest rate
is no longer controlled by the monetary authority of the country, and is defined by the
interest rate prevailing in the international markets. Thus, the central bank of each
country would lose its ability to fix the nominal interest rate in an open economy
context, regardless of its ability to control or not the money supply. On the other hand,
the relative effectiveness of monetary and fiscal policy would depend crucially on the
exchange rate regime adopted (fixed or floating).
In this paper we critically evaluate the Mundell-Fleming model3 and some of its (short
run) extensions from the perspective of the heterodox approach of the exogenous
interest rate. We will try to show that this model has some unrealistic assumptions,
such as "perfect" international capital markets and inelastic exchange rate
expectations, whetehr the latter inelasticity is relative to the financial fragility of the
balance of payments (as in the case of the fixed exchange rate regime) or relative to
the exchange rate actually observed in recent past (in the case of the floating
exchange rate regime). In the light of this critical assessment, we will show that none
of the main qualitative results of the Mundell-Fleming model on the relative
effectiveness of fiscal and monetary policies is valid and that under more realistic
assumptions the central bank has the power to set exogenously the interest rates even
in an open economy with free (but not "perfect") capital mobility. We will also see that
this power to determine the interest rate is in practice limited and can become very
asymmetric, because depending on the exchange rate regime and the conditions of
1 In the Heterodox side we have Sraffians like Pivetti (1991, 2001) and post-Keynesians like Lavoie
(2000,2001). On the Orthodox side we have John Taylor (1997, 2000), Romer (2000) among many
others.
2 Some of the few undergraduate textbooks that already include the model without the LM curve
generally do so in the context of closed economy.The open economy analysis is discussed in terms of the
Mundell-Fleming model (see Mankiw (2010) and Jones (2011)). An exception would be that of Romer
(2006). Most of the essays on pedagogy from the macroeconomic model of the new consensus without
the LM curve in Fontana & Setterfield (2009) both orthodox and unorthodox also show only the
formalization of the closed economy model.
3
For the history of the Mundell-Fleming model and its relation to the original works of Mundell and
Fleming see Darity & Young (2004).
1
the external balance of payment constraint , a given interest rate set by the central
bank can lead to various problems and may eventually become in practice
unsustainable. But, as we shall see, even in these cases the point is that in a monetary
economy there is no automatic market mechanism available to
determine the
interest rate that would be appropriate to a given economic situation and/or a given
set of policy objectives of the monetary authority.
Besides this introduction, this paper will be divided into three more sections. In Section
II, we briefly present the Mundell-Fleming model. In section III we show the results of
the model when we relax the “perfect” international capital mobility assumption. In
section IV, we present the main criticisms of the Mundell Fleming model both in the
case of fixed exchange rate and floating exchange rate regime. Concluding remarks are
made in the last section (V).
II.The Mundell-Fleming Model
II.1 IS-LM-BP with perfect capital mobility
The short-term Mundell Fleming Model is nothing more than the IS-LM model adapted
to an open economy with free capital mobility. It is assumed that nominal wages (or
alternatively the nominal price level) is exogenously given and that the level of output
can differ from the “full employment”(more generally, potential) level. The IS curve,
that show the balance in the goods market is elastic with respect to real interest rate
due to the assumption that investment is an inverse function of the interest rate and
consumption is largely induced by disposable income. In the open economy the curve
also includes net exports (exports minus imports), and these depend positively on the
real exchange rate, since a depreciation of the real exchange rate increases exports
and decreases imports. The LM curve continues to show the money market equilibria
that arise from the matching between the demand and supply of money (the latter
equal to a exogenous high powered monetary base times the given "money
multiplier").
In addition to the IS and the LM, another curve is introduced to show the conditions of
balance of payments equilibrium. The balance of payments depends both on the trade
and financial accounts. The trade account depends on exports and imports (including
nonfactor services), while the financial account depends on the nominal interest rate
differential between domestic and international bonds. A positive interest rate
differential leads to capital inflows, while a negative differential leads to an outflow of
capital. If the equality is not verified, capital flows will quickly move in the direction of
the assets with higher yields, and this will continue until the equality is somehow
restored. It is assumed that the short run mobility of international capital is both free
and “perfect” and therefore that the economy faces an infinitely elastic supply of
foreign exchange at any interest rate slightly higher than international rate. Note that
in this curve is also implicitly assumed that the market's expectation of the changes in
the nominal exchange rate change is always zero.
II. 2 The fixed Exchange rate regime
In the case of the fixed exchange rate regime , the central bank sells (and buys)
foreign exchange whenever there is a tendency to a deficit (or surplus) in the overall
2
balance of payments, losing (or accumulating) reserves , in order to maintain the
nominal exchange rate fixed. In the model, this decrease (increase) of the country's
international reserves in turn has the effect of decreasing (increasing) the monetary
base, which according to the model leads to a decrease (increase) in the money supply.
Thus, in the case of fixed exchange rate, an expansionary monetary policy produces
the following results. The increase in (exogenous) money supply decreases the
domestic interest rate below the international interest rate (plus the country’s
sovereign spread), leading to an outflow of capital and reduced net imports (since
exports are autonomous, but imports are a function of the output level). Then, the
domestic monetary authority, in order to avoid a
nominal
exchange rate
devaluation, sells international reserves at a fixed price. This leads to a decrease in
international reserves, and also in the monetary base and the money supply. When the
money supply contracts, the interest rate rises again. Note that the new equilibrium is
achieved only when the interest rate again equals the external interest rate again
and, therefore, the net effect on output is zero.
In the case of expansionary fiscal policy results are quite different. In the fixed
exchange rate regime a expansionary fiscal policy, given an exogenous money supply,
initially leads to an increase in output and domestic interest rate through the usual
the effect of the higher demand for money associated with a higher level of activity,
given an unchanged money supply. The higher domestic interest rate rises above the
external interest rate and leads to a rapid and large inflow of foreign capital. The
central bank, in order to prevent a revaluation of the nominal exchange rate is forced
to buy all of the excess of foreign exchange and to accumulate reserves. But in doing
so the monetary base and money supply expand and the domestic interest rate falls.
This process ends only when the monetary base expands enough, expanding the
money supply sufficiently to meet the increased demand generated by the fiscal
expansion at a domestic interest rate is equal to the international rate. Therefore,
monetary policy in the fixed exchange rate regime is totally ineffective in expanding
the output because the government loses control over the money supply as a
consequence of its commitment to buying and selling foreign currency. And fiscal
policy is fully effective in expanding the output because this same commitment
ultimately ensures that the traditional partial crowding out effect of private
investment that occurs in the IS-LM model of the closed economy (where the money
supply remains fixed) will not happen.
II.3 The floating exchange rate regime
In the Mundell-Fleming model with floating exchange rates, it is assumed that the
exchange rate appreciates when the domestic interest rate is above the international
level and depreciates when it is below it. If, for example, the government decides to
undertake an expansionary monetary policy by increasing the exogenous supply of
money, the interest rate decreases, thereby increasing investment and output. But
now the interest rate has fallen s below the international rate and this leads to a
devaluation. This nominal exchange rate depreciation, with fixed nominal wages
and/or prices, also leads to a real exchange rate devaluation and increased net
exports, shifting the IS curve to the right, increasing output. The increase in output,
through its effect on the demand for money, causes an increase in the interest rate.
3
But while the interest rate remains below the international rate there will be other
nominal and real exchange rate devaluations and further increases in net exports and
output. Thus output will continue to increase until the domestic interest rate rises
again to the level of the international rate. Monetary policy in the floating exchange
rate regime is thus very effective in expanding the output because it causes exchange
rate devaluations that shift the IS curve to the right.
On the other hand, in this exchange rate regime, fiscal policy loses its effectiveness
completely. An increase in government spending in principle shifts the IS curve to the
right and increases output. However, given the exogenous supply of money, the
output expansion leads to an increase in interest rates and a reduction in investment,
reducing the expansionary effect of fiscal policy. In addition to this known effect of
partial crowding out, in the Mundell-Fleming model under floating exchange rates,
domestic interest rates above the international level lead to an exchange rate
appreciation, which then reduces net exports and output. This reduction of output in
turn reduces the domestic interest rate. While domestic interest rates have not
returned to the international level, there will be other real exchange rate appreciations
and further reduction in net exports. The equilibrium will be achieved only when the
interest rate goes back to its initial level, but this will only occur when the contraction
in net exports exactly offsets the initial expansion of public spending and the level of
output also return to its initial level. Thus, fiscal policy is totally ineffective and unable
to change the level of output in the floating exchange rate regime.
III. The IS-LM-BP with “imperfect capital mobility”
III.1 Imperfect mobility?
Let us now relax the first strong assumption of “perfect capital mobility” and see what
happens to the results of the Mudell-Fleming model.
According to the extreme hypothesis of “perfect” international markets in which any
country is supposed to be a price taker in relation to the international interest rate
(plus some fixed country risk spread) and so can attract a potentially infinite amount of
foreign capital, simply by fixing an interest rate slightly higher than the international
rate.
First of all, it may be the case that the country concerned is relatively large and,
therefore, it would be unrealistic to think that the international interest rate itself is
independent of the size of its external financing operations. Secondly, even for "small"
countries - relative to the size of the global market - it could be the case that the assets
of a country are not perfect substitutes for assets of other countries and agents
require a growing interest premium to retain them in larger quantities in their
portfolios. In addition, there may be market imperfections and asymmetric
information that make the risks of loans increase when a country increases its foreign
currency liabilities. One or more of these reasons is enough to make the BP curve slope
upward in the interest-output quadrant because the international interest rate
(foreign rate plus spread) increases when the level of output increases, since higher
output brings with it higher imports and eventually growing current account deficits.
Note that the name used for this variant of the model is quite inadequate. There is
nothing "imperfect" with international capital mobility as such , which moves freely in
the short run. It is the world capital market that is not "perfect", in the sense of
4
absence of "atomism" (i.e., there are large economies) or existence of real or
perceived heterogeneous risks ("capital market imperfections").
Anyway, what interests us here is that the IS-LM-BP results with this "imperfect
mobility" of foreign capital may be quite different from the standard Mundell-Fleming
model.
In this IS-LM-BP version the crucial element is whether and when the BP is more or less
vertical than the LM, i.e., if an output expansion leads to a faster increase of the
international interest rate (which equilibrates the Balance of Payments BP) or of the
domestic interest rate (which equals the supply and demand for money LM). In the
first case, when the external interest rate rises faster than the internal rate is often
considered to be a case of “low mobility of capital and low financial integration”,
while the opposite case, when the international rate increases more slowly than the
internal rate is rationalized as a case of "high mobility of capital" or "strong financial
integration".
III.2 The fixed exchange rate regime under "imperfect mobility"
Let's see what happens with the IS-LM-BP model with "imperfect mobility", initially in
the fixed exchange rate regime. Suppose first that the BP curve is less vertical than
the LM curve. Let's start with monetary policy. An increase in money supply reduces
the domestic interest rate in relation to international and causes a capital outflow. This
reserve loss forces the central bank to buy the domestic currency and sell foreign
exchange to keep the nominal exchange rate fixed. Money supply contracts,
counteracting the expansionist policy. In this case, the result is identical to the
Mundell-Fleming model.
If now we change the hypothesis and assume that BP is more vertical than the LM and
ask again what the effect of monetary policy, the answer is the same. The initial fall in
interest rates creates a BP imbalance leading to a loss of reserves, contracting the
monetary base (and money supply). In the fixed exchange rate regime in all three cases
of perfect, high or low mobility the result is the same because in the three cases an
initial shock results in a domestic interest rates below the international rate and a
reserve loss. The conclusion that monetary policy is ineffective in the fixed exchange
rate regime is thus quite robust.
Let’s turn to the case of fiscal policy, still in the fixed exchange rate regime. Recall that
in the version with perfect capital mobility (Mundell-Fleming) fiscal policy was fully
effective and the output increased in proportion to the fiscal stimulus (i.e., the
autonomous increase in public expenditure times the traditional Keynesian multiplier
in an open economy with government). There was a partial crowding out, because
the initial rise of domestic interest rates, given the initial money supply, makes the
domestic interest rate higher than the international rate, which caused a surplus in
balance of payments, attracting foreign capital and the government was then forced to
buy foreign currencies and sell to local currency, expanding the money supply until the
domestic rate reached the international rate.
In the case of “high" but not perfect mobility the results are something close but not
identical. The fiscal expansion increases output and pushes up the domestic interest
rate, attracting international capital and the government, in order to maintain a fixed
exchange rate, buys foreign currency, which expands the money supply. This occurs
5
until the domestic interest rate is equal again to the international interest rate. But
now the international interest rate increases with the level of activity, which means
that fiscal policy will expand the economy less than in the case of perfect mobility,
because the money supply will not increase in proportion to the increase in public
spending amplified by the multiplier effect. Thus, with high capital mobility (BP less
vertical than the LM), fiscal policy is still effective but less than in the Mundell-Fleming
model.
However, in the case of "low capital mobility" the process is quite different but the
result is similar. A fiscal expansion in fixed exchange rate regime has the following
effects. The increase in government spending increases aggregate demand and output,
but now, due to "low mobility", the domestic interest rate increases less than the
international rate. This leads to a deficit in the balance of payments and a capital
outflow that forces the central bank to maintain the fixed exchange rate by buying
local currency and selling foreign currency, reducing the monetary base and the money
supply, until the equality between domestic interest rates and international rate is
restored. Again, in a slightly different way, the final equilibrium will be such that fiscal
policy will continue to have expansionary effects, but there will be a partial crowding
out effects that dampens this effect and this negative effect will be stronger the less
horizontal is the BP curve, and so, the less "perfect" is the capital mobility. At the end,
a monetary policy in the fixed exchange rate regime is still not effective and the fiscal
policy will be less effective than before, but the changes are not drastic compared to
the Mundell-Fleming model with "perfect mobility".
III.3 The floating exchange rate regime under "imperfect mobility"
It is in the case of flexible exchange rates, however, that the results change more
drastically. Suppose initially that there is "high capital mobility" and the domestic
interest rate increases faster than the international rate . A monetary expansion
causes a tendency to a balance of payments deficit that is corrected by currency
devaluations and lead to a subsequent expansion of net exports. This increase in net
exports has two effects. On the one hand it increases the level of output and the
demand for money, so that the domestic interest starts to increase. On the other
hand, the increase in net exports improves the current account balance and shifts the
BP curve itself by reducing the interest rate necessary to equilibrate the balance of
payments. In the new equilibrium, monetary policy will be expansionary although less
so than in the Mundell Fleming case, as the final effect will always be higher levels of
domestic and international interest rate , the higher is the economy’s level of activity.
On the other hand, monetary expansion in the “low capital mobility” case, when the
domestic rate increases less than the international rate, generates even more
expansionary results. An increase in money supply now generates a devaluation and
increase in net exports, which again has the dual effect of increasing aggregate
demand and output and improve the current account balance. The IS and BP curves
shift to the right until a new equilibrium level of output is reached, with an equilibrium
level of output greater than the initial level after the monetary expansion.4 With the
4
Note that the increase in output given an increase in exports is Δ X / ((1-c (1-t)) + m) and an increase in
output that equilibrates the trade account would be Δ X / m. Thus the shift of BP curve is larger than the
IS curve, ensuring the stability of the Mundell Fleming model, for all cases, since the simple Keynesian
6
increased level of activity, the domestic interest rate begins to rise again, and
eventually equals the external interest rate. In this case the result is that with "low
mobility of capital," monetary policy is powerful.
Let us now see what happens in the case of fiscal expansion, even in the regime of
flexible exchange rates with "imperfect mobility." Initially we assume "high capital
mobility", i.e., the domestic interest rate increases more than international rate. An
increase in government spending will lead to an increase in aggregate demand and
money demand, which raises the domestic interest rate. This leads to a strong
appreciation of local currency that decreases net exports, offsetting the impact of
expansionary fiscal policy. However, the decline of net exports worsens the current
account balance and also shifts the BP curve to the left, increasing the international
interest rate for a given level of activity. When the fall in economic activity caused by
the continued appreciation of the exchange reduces the domestic interest rate to the
point that equals the international rate, the exchange rate appreciation will stop and
output will cease to shrink. In the end, the crowding out effect will be less than
complete and fiscal policy will have had some expansive effect, even with flexible
exchange rates.
Finally, we come to the case where there would be "low-mobility" of capital and the
domestic interest rate increases more slowly than international rate when the level of
output increases. In this case, fiscal policy with flexible exchange rates would have the
following effects. An initial increase in output increases the domestic interest rate,
given the money supply. Given the assumption that the domestic interest rate
increases more than the international rate, the economy is now below (not above) the
BP curve. Devaluations occur and increase net exports. Thus, not only the fiscal
expansion is not neutralized, but it still indirectly causes an expansion of net exports
and therefore a further increase in the levels of output. In addition, the increase in
net exports causes a shift of the BP curve to the right, because the current account
balance improves. This process ends only when the domestic interest rate rises so
much that equals the international interest rate. The model is stable because the
expansion of the IS increases the domestic interest rate (given the money supply) and
shifts the BP curve to the right, reducing the international interest rate. In the end,
fiscal policy in the flexible exchange rate regime may be more expansionary than in the
fixed exchange rate regime, even more than the fixed exchange rate with perfect
mobility, end this result depends only on the assumption that capital mobility is
sufficiently "imperfect."
What could we conclude in this section through this complicated sequence of
arguments about the IS-LM-BP with "imperfect mobility" and the Mundell-Fleming
model? Some authors such as Dunn jr. & Mutti (2005, p.489) conclude that fiscal policy
is very effective with fixed exchange rates and high or perfect capital mobility and
under flexible exchange rates and low capital mobility. Monetary policy in turn never
works in the fixed exchange rate regime, but in the floating exchange rate regime it
works well with perfect mobility, and less well with both high and low mobility of
capital.5
models the usual stability condition (c (1-t) <1) is met. For an analysis of the formal conditions for
stability of the IS-LM-BP, see Gandolfo (2002).
5
Most textbooks omit the case of monetary and fiscal policy under low capital mobility. Some examples
of books that explore some of these effects are Dunn jr. & Mutti (2005, p.482-3), Cysne Simonsen (1989,
7
Let's see what we can conclude taking a more critical view of these models. As noted
above, the idea of an upward sloping BP curve is reasonable, although in fact it has
nothing to do with “low capital mobility”. However, the notion that appears in the ISLM-BP using the BP curve slope in relation to the LM curve slope as a measure of the
degree of international mobility of capital is basically meaningless, because the slope
of the LM curve reflects only parameters such as the interest and income elasticity of
demand for domestic currency. Using these slopes as a measure of relative mobility of
foreign capital implies absurd results. In fact , everything else being constant, if there is
a country A where by law capital is as free to move abroad as a country B but the
parameter of the demand for money for the transaction motive of country A happens
to be greater than that of country B for some reason, then we must conclude that
country A will have necessarily a “lower degree of international capital mobility” than
country B.
But in fact, we should not worry much about the different results of the IS-LM-BP
models regarding the appropriate of economic policy mix in each exchange rate
regime and “degree of capital mobility”, simply because both the position and the
slope of the LM are irrelevant, because in practice central banks set directly the basic
domestic interest rates thus the "demand for money" has no direct effect on these
rates. As the LM curve has been in practice abandoned (or to be more precise, made
redundant) even in the New Consensus Macroeconomics, we should think of
discarding it also in the open economy analysis and thus forget the IS-LM-BP model
with its unnecessary complications.
IV. Critical evaluation to the Mundell-Fleming model
IV.1 Introduction
In the previous section we saw that the policy mix conclusions of the Mundell-Fleming
model conclusions are not completely robust when we relax the extreme hypothesis
of perfect foreign capital mobility (horizontal BP curve). However, this could turn out
not to be such a big problem in practice, to the extent that one could argue that the
MF model with its s extreme assumptions and sharp results could still be admittedly
a highly simplified but still the most relevant model for a first approximation to the
analysis of the problems of many economies in the modern world that are
characterized by a very high degree of financial openness. In this section we will try to
argue that this is not the case and that the model is not a good characterization of the
simplified short-term behavior of a financially open economy, and we will make some
simple suggestions on how to treat these issues with a bit of realism.6
p.351) and Froyen (2003). The reason for the lack of concern about the results of BP curve more steeper
than the LM curve, as Froyen said is that "Although the result in Figure 7.21 [BP steeper than LM] is
possible in theory, most economists believe that the result of Figure 6.21 [BP flatter than LM] is more
likely. They find that an expansionary fiscal policy will reduce the exchange rate (increase the value of
domestic currency in relation to other countries). This belief stems from the idea that there is a
relatively high degree of international capital mobility, which means that BP is relatively flat, and
therefore likely to be flatter than the LM curve (...) [emphasis added] " (Froyen, p. 591)..
6
Godley and Lavoie (2004) introduce the exogenous interest rate in a model that they consider "small",
with "only 32 equations" and all the international spillover effects on flows and stocks, which do not
8
IV.2 The fixed exchange rate regime
a) Increasing Spreads and Credit Rationing
The first issue to be criticized is the idea of an unlimited supply of foreign capital at a
domestic rate of interest slightly higher than the international interest rate. Even if the
international rate includes a spread that reflects the risk and / or illiquid liabilities in
foreign currency of a country, it seems clear that even with free capital mobility (in the
sense that there are no obstacles to entry and exit of capital flows from abroad in the
economy) and even in the short term, it is not realistic to think that any current
account deficit of any magnitude can always be financed at a constant interest rate.7
In addition to increasing risk, which seems inevitable beyond the point at which the
expansion of the economy leads to rising current account deficits, it is possible to state
that beyond a certain limit begins foreign credit rationing is bound to occur and the
country will not be able to obtain more foreign currency, regardless how high is the
domestic interest rate. In other words, the BP curve, in addition to being positively
sloped, becomes vertical at a certain point. This relationship is shown in Figure 1
below, where i is the domestic nominal interest rate, i* the international interest rate,
and Y the output level.
help to separate the effect of this hypothesis from all the others that become necessary to make the
model tractable. Recently, Lavoie and Daigle (2011) introduced also the possibility of currency
speculation and endogenous expectations in a simulation model and seem to confirm its destabilizing
role in a flexible exchange rate regime. Ferrari-Filho (1994) criticizes the Mundell-Fleming model from
the post-Keynesian perspective of "fundamental uncertainty", quite different from that adopted here.
7
An important point emphasized by Smithin (2002-03) and Lavoie (2002-03) is that what is most
important is not the assumption of perfect capital mobility, but that of perfect substitutability between
assets in the world. Without the latter, there would be an infinite input of capital when a country raises
its interest rate above the international (plus the effect of risk and the expectation of devaluation), even
if it was fixed for a long period of time. Thus, the hypothesis of imperfect substitutability between assets
globally is enough to ensure that there will be no endless flow of capital into the country who fix an
interest rate higher than the world. This means that if there is a hierarchy of international currencies
according, for example, how deep and liquid are international assets markets denominated in that
currency, at least part of the spreads charged by international investors do not merely reflect the
default risk assessment in the external accounts of these countries but also the relative degrees of
international liquidity of the currency of those countries.
9
Figure 1 –BP curve with increasing spreads and credit rationing
i
Y
a) Expectations of Devaluation
Besides the increased risk and the possibility of international credit rationing, in the
specific case of the fixed exchange rate regime it is also necessary to take into account
that, as the economy expands and incur in increasing current account deficits with
potential losses of reserves, it is likely that, despite the fact that the exchange rate
regime is fixed, in financial markets inside and outside the country's growing doubts
about the sustainability of exchange rate parity chosen by the government may arise.
These doubts tend to turn quickly into expectations of a arge exchange rate
devaluation. These expectations of a change in the nominal exchange rate will move
the whole BP curve upwards and may generate a process of continuous loss of
reserves. Such a process tends to be offset by the central bank directly increasing
domestic interest rates to a level that compensates the expectations of devaluation (or
what is less common these days, imposing capital controls or by the taxation of capital
outflow).8
So one of the most problematic assumptions of the Mundell-Fleming model is an
implicit assumption of full credibility of the fixed exchange rate regime, because only
then the expected depreciation would be zero and uncorrelated with the loss of
reserves and the size of the current account deficits and external debt of the country.9
b) Exogenous interest rate and non-sterilized interventions
8
Santos (2005) makes similar criticisms to the Mundell-Fleming model but unfortunately related the
notion of limits to the attraction of foreign capital to the orthodox argument for the portfolio
distribution between different assets by Tsiang (1975), which depends on an implicit and totally
unrealistic assumption a of world exogenous money supply. On the other hand Santos(2005) assumes
that the BP curve becomes vertical at a critical level of the external interest rate equal to the rate of
growth of exports because beyond that foreign debts become unsustainable. But in reality the curve
may become vertical before that because of the problem of external liquidity, i.e., if the level of short
term external debt (not total external debt) becomes very high relative to foreign exchange reserves (on
that see Medeiros & Serrano(2006).
9
For an analysis about the collapse of the Brazilian managed exchange rate regime in 1999 as a process
of accumulation of current account deficits, growing risk and endogenous expectations of devaluation
see Medeiros & Serrano (2006).
10
Another crucial and totally unrealistic point in the Mundell-Fleming fixed exchange
rate regime model was noted by Lavoie (2002, 2005) and concerns the transmission
channel from foreign exchange reserves to domestic interest rate. In the MundellFleming model a balance of payments surplus increases the monetary base, and this in
turn increases the money supply, shifts the LM curve, reducing the economy's
domestic interest rate. In this case we say that foreign exchange interventions are nonsterilized.
The problem is that if we assume, with a minimum of realism, that in practice the
central bank operates by setting the basic nominal interest rate, this mechanism no
longer makes sense. In this context, in many cases, the entry of foreign capital that
caused the increase in reserves will go straight to the purchase of government bonds
in the country (or be invested in local private banks and funds that buy and hold these
bonds) which has set the domestic interest rate at an attractive level , above the
international rate. In this case the "sterilization" of the increase in reserves will be
voluntary, full and automatic as the whole of the increase in foreign reserves is
offset directly by an equal increase in domestic public debt and there is no increase in
the monetary base (and of course also no increase in the money supply M1), and the
domestic interest rate remains unchanged.
In other cases, the increase in reserves may have been the result of an increase in
exports or other financial flows or remittances to the country. In these cases, in fact
the monetary base increases when the first foreign currency are exchanged for
domestic currency at the fixed exchange rate. However, there is no reason to suppose
that it affects the domestic interest rate previously set the central bank. In the first
moment, the private bank reserves actually increase, but there is no reason to suppose
that the banks are forced to lend these additional resources if nothing occurred to
increase the number of customers considered creditworthy by banks at the current
level domestic interest rates (plus banking spread). As in a modern capitalist economy
banks are not forced to lend, the only direct impact of the initial monetary base
increase is a drop of the money “multiplier" (monetary base - money supply ratio). On
the other hand, the basic domestic interest rate, fixed by central bank also has no
reason to be altered and will remain constant. In fact, sooner or later what is most
likely to occur is that banks do not want to retain these extra idle and unpaid reserves,
and in the end will invest these excess bank reserves in government securities. As the
central bank, when setting the basic interest rate, has of course agreed to sell bonds
at the exact amount that is demanded, this leads to an increase in public debt and
sterilization of the total increase in foreign exchange reserves, because the increase in
public debt reduces again the monetary base to its previous level.
It could be also be the case that those who sold foreign currency and bought domestic
currency in fact wish to keep in domestic cash a part or all of the initial increase in
foreign reserves increase. In this case, which would include, for
example, foreign
tourists in need of local bills and coins for small expenses or a foreign owner of a
supermarket chain that has sent money from abroad to its subsidiary in Brazil and
needs cash and coins to give small change in their stores. In this case the increased the
newly issued domestic money will not be deposited in banks and will not be applied
in government bonds. There will be no sterilization and the monetary base will
increase. However, even here there is no reason for a change in this economy's
domestic interest rate, because the increased local currency held by the “public” was
11
entirely desired by the public and there were no unwanted excess local currency that
could affect the interest rate ( the “money multiplier” falls). Note that if the tourists
later, the manager of the supermarket or any member of the public who receive the
notes and coins corresponding to the base increase resolve to get rid of these, they will
directly, or through funds, to invest these funds in government bonds, which will be
issued to maintain the basic interest rate and so finally these resources will also be
"sterilized." Note that nothing needs to be done by the central bank specifically for the
purpose of sterilization. If we just assume that the central bank wants to maintain the
fixed exchange rate and the benchmark interest rate unchanged , the sterilization will
necessarily occur in the exact amount that is required to achieve these two objectives
simultaneously.
As Lavoie said “Neutralization arose either automatically, at the initiative of the private
sector, or naturally, as a result of the normal behaviour of the central bank to sustain
the payment system” (Lavoie, 2005 p.29).
So he correctly concludes that “fixed exchange rates did not prevent central Banks
from setting interest rates, while money creation was still demand-led” (2005, p. 29),
that is, no matter what may happen to the monetary base, the amount of money
created in a system of fixed exchange rate is determined not by the balance of
payments but for the demand for bank loans of creditworthy clients given the current
basic interest and bank spreads, just as in closed economy. The sterilized intervention
is almost always the rule and not the exception to the rules of the game.
c) Capital inflows, changes in reserves e and credit expansion in fixed Exchange
rate regimes
Note that in a fixed exchange rate regime, unlike the automatic adjustment assumed in
Mundell Fleming model, neither the interest rate or the amount of domestic credit is
directly affected by a surplus or deficit in the balance of payments. So capital inflows
do not by themselves directly generate credit booms even in a currency board regime
where the printing money has supposedly to be fully backed for foreign currency.
What actually happens in countries that adopted this extreme system is that domestic
banks do lend the amount of local currency to customers whom the banks consider
solvent, given the basic interest rate and banking spreads. In this case, an increase in
bank loans within the country tends to create the usual need to increase bank
reserves, a need that is fulfilled to the system as a whole, either through increased
central bank lending to the private banking sector, or by the sale of government
securities from banks to the central bank. In both cases the monetary base expands
endogenously as a consequence of endogenous expansion of credit and the “money
supply”.
Of course this increase in the monetary base decreases the degree of coverage of the
exchange system, i.e., the monetary base becomes greater than the level of foreign
reserves. This forces the government of the country to try to attract more foreign
funds to increase foreign reserves and restore the reserve coverage of the currency
board. The end result of all this will be a strong increase in external public debt (in
foreign currency) with the international private sector (and / or the metropolitan
government in the case of colonies that have used this scheme) which becomes
necessary even if the domestic banking system lend to activities that use very few
12
imports and thus require little or no foreign currency such as construction, for
example. Currency board schemes, especially those that depend on private external
credit tend to be short lived due to a combined effect of strong appreciation of the real
exchange rate due to domestic inflation remaining above international levels and due
to the effects described in items a) and b) above. But the point is that even in this
kind of regime the domestic interest rate is exogenous and the money supply
endogenous, because local banks are not forced to lend to those who they do not
want to and the central bank does not lose the power to determine the basic interest
rate in domestic currency.10
This does not mean that we do not often observe an empirical relationship between
capital inflows, reserve accumulation and the expansion of credit and effective
demand of the local economy. But the correct chain of causality comes from more
favorable external interest rates and credit terms in relation to domestic credit and
interest rate conditions, inducing local banks or businesses and workers to take more
credit to finance their decisions to increase spending on productive investment,
residential investment and durable consumption and / or buying existing assets. The
accumulation of reserves is only the inevitable consequence of the exchange rate has
been kept fixed11 in this process if the foreign capital net inflows is higher than the
expansion of imports induced by this increase in credit and spending. And a possible
change in monetary base is a mere consequence of, as we saw above, the public and
the banks needs of holding more money in liquid form and by no means impacts the
domestic interest rate or the actual expansion of domestic credit.
d) Asymmetries
Since the basic endogenous mechanism of Mundell-Fleming model with fixed
exchange rate - the foreign reserves – domestic base nominal interest rate connection
- does not work, it is clear that in the fixed exchange rate regime there is no good
reason to think that fiscal policy would be more or less effective as compared to what
would occur in a closed economy in terms of expanding or contracting aggregate
demand and output.
What is relevant in the fixed exchange rate regime with free capital mobility are two
strong asymmetries. The first is between conditions where the economy is above or
below the BP curve , i.e., accumulating or losing reserves. In practice there is no limit
to how much a country can accumulate international reserves. In fact, the only cost of
this process is the expansion of public debt that comes from the partial or complete
sterilization of the foreign reserves that is required to maintain the basic interest rate
set by the Central Bank. This cost, which is the difference between the domestic
interest rate and foreign interest rate paid in the country's reserves can lead to
extensive government transfers of income to holders of public debt. However, it is
unlikely that such transfers, even when large, should have a significant impact on
aggregate demand , because it is highly unlikely that the banks and rich agents who
10
Even the excellent work and Manoelito and Calcagno (2003) on the period of convertibility in
Argentina incurs in this misinterpretation. For a brief analysis of the operation of schemes currency
board approach the light of the exogenous interest rate see Serrano (2003).
11
In fact it is sufficient that the exchange is not fully flexible (“dirty floating”) to generate some increase
in reserves under these conditions.
13
hold these resources would greatly increase their spending as a result of this increase
in domestic public debt and its service.
On the other hand, the popular belief that an increase in domestic debt affects the risk
spread to the country's external loans in foreign currency and, thus, the BP curve, also
seems incorrect. As a matter of logic, governments that issue their own currency
cannot be forced by the market or anyone else to default on its domestic debt
denominated and paid in its own currency, so it is impossible that the government
faces a lack of money to pay its internal public debt.12 Moreover, the idea that,
collectively, the international investors who are creditors of the country in
international currency (usually U.S. dollars) would be concerned with any kind of
indicators of the local economy that do not affect either the country´s nor the
government´s cash flow and ability to pay its commitments in foreign currencies also
simply makes no sense.
The correlation sometimes observed between basic interest rates and public debt,
therefore, has the causality coming from the interest rate fixed by the Central Bank to
domestic public debt, and not the opposite. And as the central bank tends to fix its
domestic interest rate above the foreign rate (the rate plus the external spread outside
the country), the sequence of causality that makes sense begins in the country’s
sovereign spread, which depend on international market conditions and the overall
credit and market assessment of the liquidity and sustainability of the external
accounts of the particular country. These external changes in the spread tend to affect
the domestic interest rate required to maintain the fixed exchange rate, and
subsequently this increase in interest rates impacts the domestic country’s debt.
The above discussion shows that there is no element that forces the central bank to
stop accumulating reserves at a fixed exchange rate regime, contrary to the MundellFleming model. On the other hand, no matter how large is the initial stock of reserves
and the initial access to credit in foreign currency of a country, usually the reserves are
finite and can, in a context of free international capital mobility, fall very rapidly, a
process which tends to be magnified with the emergence of expectations of an
eventual devaluation of the exchange rate, which leads to further outflows of
speculative capital. These movements can rapidly make the fixed exchange rate regime
impossible to maintain. Thus, although in principle the domestic interest rate may be
fixed by the monetary authorities both above and below the levels specified by the BP
curve, there is a strong asymmetry between positions above and below this curve.
Domestic interest rates fixed below the curve can lead to unsustainable situations, and
to an upward shift in the BP curve for two complementary reasons. The first would be
to increase the external spread for a given level of activity and current account deficit,
due to the strong loss of reserves. And second, the emergence and subsequent
12
About this Tcherneva (2011) transcribes a long quote from Orthodox Michael Woodford: “A subtler
question is whether it makes sense to suppose that actual market institutions do not actually impose a
constraint … upon governments (whether logically necessary or not), given that we believe that they
impose such borrowing limits upon households and firms. The best answer to this question, I believe, is
to note that a government that issues debt denominated in its own currency is in a different situation
than from that of private borrowers, in that its debt is a promise only to deliver more of its own
liabilities. (A Treasury bond is simply a promise to pay dollars at various future dates, but these dollars
are simply additional government liabilities, that happen to be non-interest-earning.) There is thus no
possible doubt about the government’s technical ability to deliver what it has promised…” (Woodford
2000, p. 32)
14
increase in the expected devaluation of exchange rate by the market. Asymmetrically,
domestic interest rates above the curve indicated by initial BP, despite their fiscal costs
can be kept for long periods of time without creating situations that require its rapid
reversal.
The other important asymmetry in fixed exchange rate regimes with free capital
mobility would be between contractionary and expansionary policies. In situations
where the economy is still in the region above the curve BP any expansion, be it a fiscal
or monetary (i.e., the shift to a lower interest rate but still above the international
rate) leads to an increase in imports and the loss of reserves that ,from some point on,
will increase both the current account deficit and the ratio of short-term external debt
over reserves and thus begin to worsen both indicators of external debt sustainability
and of the economy's external liquidity. On the other hand, asymmetrically
contractionary monetary and fiscal policies (even assuming that the economy is
initially at a point above the curve BP) have the opposite effect of reducing the current
account deficit and increasing the stock of reserves.
These are the problems and asymmetries that affect the fixed exchange rate regimes
in the exogenous interest rate approach. The supposed greater effectiveness of fiscal
policy in relation to monetary policy in the Mundell-Fleming model does not occur,
even less the supposed impossibility of implementing a monetary policy, which comes
from a presumed "trilemma" between free capital mobility, fixed exchange rate and
monetary policy autonomy that appears in literature inspired by this model. At least
for contracting effective demand, a monetary policy of raising domestic interest works
without any problems in a fixed exchange rate regime and free capital mobility (even
in a currency board system).
IV.3 Floating Exchange Rates
Many of the criticisms we made to the Mundell Fleming model with fixed exchange
rates will also apply to the floating exchange rate version, particularly with regard to
the possibility of increasing risk spreads and credit rationing beyond a certain point.
However, some of the problems of the model appear in modified form in the context
of a floating exchange rate regime.
a) Exchange rate and general price level
Even assuming that initially nominal wages and general price level is constant, a
devaluation of the exchange rate will increase the general price level by increasing
domestic prices of all tradable goods that are priced in foreign currency outside the
country in certain market internationally. On the other hand, an appreciation of the
exchange will have the opposite effect of generating a reduction in the general price
level by decreasing the level of prices of tradable goods in local currency. Weeks
(2009) draws attention to the fact that the Mundell-Fleming model in the case of
flexible exchange rate incoherently ignores these effects on the general price level in
the economy. Suppose that there is no wage resistance, i.e., subsequent increases in
nominal wages in an attempt to restore the purchasing power of wages and also there
is no other kind of reaction from other groups of agents or any type of formal or
informal inflation indexed contracts in the economy, so that a process of permanent
15
inflation (or deflation) does not happen. Yet, the inevitable initial increase in the
general level of prices after a nominal devaluation (or reduction after the nominal
exchange rate appreciation) will have impacts on the model results. As Weeks (2009)
shown, taking into account this effect, the expansive monetary policy would lose its
power, and fiscal policy becomes more effective. In the case of monetary policy, an
expansion of money supply leads to a nominal devaluation that, by causing an increase
in general price level, reduces in part the actual increase in the real money supply
which is what determines the domestic interest rate of the economy. Thus, the higher
the proportion of tradable goods, the lower the expansionary impact of a nominal
increase in the money supply. On the other hand, an expansionary fiscal policy with
constant nominal money supply leads initially to increased output and a higher
nominal interest rate, which causes an exchange rate appreciation which reduces net
exports. The nominal exchange rate appreciation, however, will reduce the general
level of prices and thus cause an increase in real money supply and an increase in
investment. Thus, following Weeks (2009), monetary policy loses its effectiveness and
fiscal policy will gain some effectiveness, the greater the proportion of tradable goods
in output and the higher is their impact on the general price level. And we may add
that if a devaluation triggers a process of permanent inflation, monetary policy lose
completely their effectiveness and the fiscal policy gain overall effectiveness in the
longer run, if we still maintain the questionable hypothesis of an exogenous money.
b) Exogenous Interest Rate
As we saw in Section II above, the endogenous mechanism of automatic adjustment of
the Mundell-Fleming flexible exchange rate regime is based on the idea that a
domestic interest rate below the international (plus sovereign spread) rate will lead to
a tendency of capital outflows that will cause a real devaluation, large enough so that
this outflow is offset by an increase in net exports. The domestic interest rate then will
again rise to the international level because the devaluation, as it expands net
exports, also expands aggregate demand and output, and with it the demand for
money, which increases domestic interest rates, if the money supply is exogenous.
Faced with an initial situation in which domestic interest rates are below the
international rate, capital outflows, currency devaluation and the increase in net
exports will continue to occur until the output is increased just enough to increase the
domestic interest rate back to international level. The same kind of automatic process
to occur symmetrically if, for some reason, the domestic interest rate is above the
international rate. In this case there would be capital inflows, exchange rate
appreciation that continues until the net exports and the output fall enough so that
the demand for money is reduced to the point that the interest rate fell back to the
international level. That is this very peculiar mechanism of automatic adjustment of
domestic interest rate to the international rate which upholds the idea that monetary
policy would have expansionary effects in flexible exchange rate regime, as increases
in money supply increase permanently net exports and output. This is also the
mechanism behind the idea that fiscal policy is impotent, as a fiscal expansion, given
the exogenous supply of money, would cause exchange rate appreciation and
reduction of net exports until output, the demand for money and interest rate return
to its initial level.
16
If we assume more realistically the domestic basic rate of interest is exogenous and
the money supply is endogenous it is evident that this automatic mechanism stops
working. With the exogenous interest rate below the international rate, continuous
devaluations occur and even if the economy expands, the interest rate will not
increase because the amount of money expands in line with the increasing activity
levels (and even prices) of the economy. Thus, a fiscal expansion, with the domestic
interest initially equal to the international level will not be offset by a fall in net
exports, since there is no reason to suppose a tendency towards currency
appreciation. On the other hand, a monetary expansion (fixed interest rate below the
international level) would lead to endless and continuous devaluations, because the
exogenous domestic interest rates will not return to the level of the international rate.
The same process would occur, also without limit, in the case of a restrictive
monetary policy. There would be a process of continuous exchange rate appreciation,
a cumulative process that could not be stopped endogenously but only by an
autonomous decision of the Central Bank to change the domestic interest rate again.
This shows that none of the basic results of the Mundell Fleming model with flexible
exchange rate survives in a more realistic context of exogenous interest rates.
c)Are Exchange rate devaluations always expansionist?
Another limitation of the Mundell-Fleming model with flexible exchange rates is that it
assumes that exchange rate devaluations are always expansionary and currency
appreciation are always contractionary. A real devaluation usually reduces real wages,
given nominal wages, because increases in tradable goods prices (and / or inputs),
transfer real income from the workers to the business and the “rest of the world”
sectors. If we consider that the workers’ propensity to consume is higher than that of
capitalists, this redistribution reduces aggregate consumption. Accordingly, a a real
exchange rate appreciation will generally have the opposite effect of increasing the
aggregate consumption.
In addition, in certain situations, private companies and/or banks of a country can have
a lot of debt in foreign currency. In this case a large real depreciation of local currency
can lead to a financial crisis for companies that will have to cut spending and/or banks
that may be forced to contract credit. Again the effect of a devaluation on real
consumption and investment can be quite negative, while an appreciation would have
the opposite effect of reducing the weight of private external debt in local currency.
Obviously this last financial effect will occur in reverse if the economy in question has a
private sector that is a net creditor in foreign currency. In this case, it is the real
appreciation that creates financial difficulties by reducing the value of foreign assets of
banks and local businesses.
The Mundell-Fleming model ignores all these possible effects, but nothing guarantees
that the positive effect of a real depreciation on net exports will predominate.
There is also the possibility that a devaluation is partly eroded by a possible increase
in inflation from further increases in nominal wages in response to actual initial losses,
the "real wage resistance". The stronger this effect will lower the real exchange rate
depreciation associated with a nominal devaluation, and thus lower will be the
expansion of net exports. On the other hand, the s contractionary effect on
17
consumption of workers and the adverse financial effects, in the case that companies
and banks are net borrowers in foreign currencies would also be lower. All these
possibilities show that, contrary to the postulates the Mundell-Fleming model, not all
real devaluations expand the output13, even if and when net exports expand. 14
d) Endogenous expectations of change in Exchange rates
Last, but not least, we must criticize the flexible exchange rate regime Mundell
Fleming model in what the striking lack of realism of the assumption of expected
level of the exchange rate in subsequent periods is exogenous15 and identical to the
initial exchange rate in the more relevant short run version of the model. As noted
above, this hypothesis in the case of fixed exchange rate regime is totally unrealistic
and assumes a priori and implicitly that the fixed exchange rate regime has full
credibility even though there may be huge losses of reserves and / or the current
account deficits may be growing without limit. In the case of a flexible exchange rate
regime the static expectations hypothesis is even less plausible because in the model
the current exchange rate is now an endogenous variable that will necessarily change
vary over time.
Such hypothesis could only make sense if the adjustment of net exports and current
account was faster than the adjustment of the capital account. In this case, even with a
big shock, only a once for all depreciation or appreciation would be necessary to
restore the balance of payments equilibrium of the economy. Only in this case, there
would be no time, nor need to revise the expectations of change in the exchange rate
after a shock that altered the current exchange rate that equilibrates the BP, because
the imbalance would already have been eliminated.
Perhaps (with a lot of good will) this hypothesis about the relative speed of
adjustment of the capital account and current account could even make some sense
for developed countries at the time that the Mundell-Fleming model was formulated,
in the days when people discussed if it would be better to replace the Bretton Woods
system with fixed or flexible exchange rate (i.e., in the 1960s). At that time there
were strong short-run capital controls in many countries and it was still plausible to
13
Although Kalecki already referred to these problems in the 1930s the literature on the possibility of
contractionary exchange rate devaluations has grown since Krugman and Taylor (1976). See the recent
survey by J. Frenkel (2011). Many empirical studies on the subject find a contractionary effect of
devaluation in the short run, followed by an expansionary effect later. Most likely what happens is that
the direct effect of devaluation on the effective demand and output are most often contractionary but
on the other hand, by often loosening the external constraints on growth, make possible, though not
determine, the pursuit of more expansionary macroeconomic policies later.
14
Note that we are not talking here about the known issue that the actual net exports will only increase
with a real devaluation if the so-called Marshall-Lerner conditions are met, i.e., the sum of absolute
values of the real exchange rate elasticity of imports and exports is greater than one, which also may not
always occur in practice.
15
In the long-run versions of the IS-LM-BP, this case it is even more unrealistic since the expected
change in exchange rate evolve according to the expected differential between domestic inflation and
international inflation, in order to ensure the validity of purchasing power parity in the long run. We are
not concerned here of the critique of this extreme view , as we focus here on the most relevant model,
which is the Keynesian or "short run" version,where the only relevant hypothesis is that the expected
future exchange rate is exogenous. For the critical conditions of long-run parity of purchasing power
parity and real interest rates, see Lavoie (2000).
18
imagine a world where capital account movements were dominated by long-run
capital in terms of both loans (official and to a lesser extent private) and foreign direct
investment . However, in current times of high mobility of short-run capital flows this
idea makes no sense at all.
But if , in reality, capital flows are fast and changes in net exports subsequent to an
change in the exchange rate are slow, then it is clear that any imbalance in the model
will not be corrected by a single exchange rate appreciation or depreciation but, at
best, by a sequence of appreciations or depreciations over time.
But this makes it impossible to sustain the idea that the expected exchange rate in a
future period is not affected at all by the exchange rate that actually occurred in the
recent period. But if we suppose, with a little realism, that expectations are minimally
elastic with respect to the exchange rate actually observed, the short-run Mundell
Fleming model can become quite unstable. Recalling that the equilibrium position of
the balance of payments curve (either perfect or imperfect mobility, large or small)
depends on the equivalence between the domestic interest rate and international
(including the spread of external risk) plus the expected change in exchange rate, each
time, for example, that domestic interest rates fall below the international and the
current exchange rate begins to depreciate, the spot exchange rate expected for a
subsequent period will also change in the same direction, increasing the rate of change
of expected exchange rate and therefore shifting the BP curve upwards, which
increases the initial divergence between the domestic interest rate and international
rate and causes an acceleration in devaluation that will turn the switch back
expectations of change exchange, and so on. The same effects occur symmetrically in
the case of domestic interest rates above the rate outside: there will be a tendency to
a continuous appreciation of the exchange rate and the BP curve will shift down
continuously as expectations of currency appreciation increase16. This problem can
be seen in Figure 2 below:
Figure 2 – Exchange rate devaluation and an expectation cumulative process of
devaluation
i
Y
16
Note that our argument here depends only on the assumption that the “elasticity of exchange rate
expectations” in the sense of Hicks (i.e., the expected change in the exchange rate from a change in the
actual exchange rate) is positive, even if it is lower than one.
19
e) There is no such thing as a free floating regime
If we combine the more realistic assumptions of exogenous interest rate and elastic
expectations of exchange rate the result is that a floating exchange rate regime
becomes violently unstable subject to a cumulative process of exchange rate
appreciation or devaluations.
From this we can draw two conclusions. First, the Mundell-Fleming flexible exchange
rate with its implausible idea that any deficit in capital account will be automatically
and quickly eliminated by the instant effect of changes of exchange rates on net
exports and the current account is not an useful tool of economic policy analysis even
as a first approximation in a high level of abstraction.
And secondly, from this critique of the Mundell-Fleming model we can better
understand why in the real world with high international capital mobility there is
simply no free floating exchange rate regime in the short-run without intervention by
the central bank, whether acting directly in foreign exchange market to try to stabilize
the current and expected exchange rate, or a little more discretely varying the
domestic interest rate in order to control minimally the direction and speed of
evolution of the nominal exchange rate at each time according to its main policy
objectives.
Note that this is not just the well-known and important case of "fear of floating"
whereby central banks of net foreign debtor countries (public and / or private) and
with high inflation try to moderate the trends of currency devaluation and central
banks of net foreign creditor nations try to moderate trends of exchange rate
appreciation so as not to lose market share in world trade and avoid major losses in
the local value of assets of the domestic financial sector. It is something much more
general serious, in that, with free capital mobility, markets could go either toward
cumulative appreciations or depreciations. This would be the "fear that the
fluctuations of exchange rate will never stop" by themselves if the central bank does
not do something to stabilize exchange rate expectations.
V. Final remarks
This critical view of the Mundell-Fleming model allows us to draw three main
conclusions. The first is that from the analytical point of view, the hypothesis of
interest rate as determined by an exogenous supply of money is unrealistic in the
closed economy and as inadequate or perhaps even more in an open economy,
implying that non-sterilized interventions in practice are the exception rather than the
"rule of the game."
The second is that the central problem with regard to economic policy is not the
exchange rate regime as such, nor the choice between monetary or fiscal policy as the
most effective tool. The central problem is the issue of free short run capital mobility.
It is this free mobility that makes it difficult to maintain a parity in the fixed exchange
rate regime, because once we have widespread expectations of devaluation it is
difficult to avoid an endogenous worsening in the BP curve, and the consequent need
20
to increase domestic interest rates and/or cut government spending to lower imports.
It is free capital mobility which also makes a truly “clean” floating regime impossible,
because the endogeneity of exchange rate expectations can lead to a continued
process of appreciation or depreciation of the nominal exchange rate, forcing the
government to intervene either directly buying or selling foreign currency directly or
indirectly through the manipulation of interest rates. So is the free mobility of capital
that takes away degrees of freedom of expansionist policies, whether fiscal or
monetary. On the other hand, because the asymmetry of the balance of payments
constraint, unfortunately there is no automatic mechanism to prevent or correct an
interest rate that is too high in comparison with international rates or excessively
contractionary fiscal policies.
Finally, the above analysis confirms the idea that even in the open economy with free
short run capital mobility, the short-term basic interest rate is exogenous in the sense
of being operationally an economic policy variable, directly controlled by the central
bank. This, due to the problems and asymmetries discussed above, does mean that in
principle the central bank may fix the interest rate at any desired level. In practice, this
power is limited by the different consequences on the balance of payments situation
and the evolution of the exchange rate of a particular level of interest rate chosen by
the monetary authorities. The fact that the central bank sets the interest rate, but on
the other hand the interest rate fixed under free capital mobility are not always
sustainable for a long period makes it clear that we are not saying that the central
bank has absolute power but that, realistically, there is simply no automatic
endogenous mechanism of adjustment in the interest- exchange rates and balance of
payments system, which makes it inevitable that the exchange rate regime is always
partially administered with both buying and selling of currencies by central bank or via
manipulation of the domestic interest rate, and makes useful some measures of capital
controls (particularly short-run foreign capital inflows taxation), so that they expand
the degrees of freedom for macroeconomic policy, both monetary and fiscal.
Bibliografia
Calcagno, A. and Manuelito, S. Argentine convertibility: is it a relevant precedent for
the dollarization process in Ecuador? Estudios estadísticos y prospectivos,No. 15 ,
CEPAL, 2001
Dunn Jr, R. ; Mutti, J. International Economics 5 ed. Routledge, 2005.
Ferrari-Filho, F. Os modelos "keynesianos" de economia aberta e as suas
inconsistências keynesianas. Texto Didático n° 01, UFRGS,1994
Fontana G. & Setterfield, Macroeconomic Theory and Macroeconomic Pedagogy
Palgrave Macmillan, 2009
Frankel, J. A. Monetary Policy in Emerging Markets: A Survey. HKS Faculty Research
Working Paper Series RWP11-003, January 2011.
Froyen, R. Macroeconomia, 3. Ed, Saraiva, 2003.
Gandolfo, G. International Finance and Open-Economy Macroeconomics, Springer
Verlag Berlin: Heidelberg. 2002
21
Godley, W.; Lavoie, M. Simple open economy macro with comprehensive accounting a radical alternative to the Mundell Fleming model , working paper n. 15, Centre for
Financial Analysis & Policy, Cambridge University, 2004
Jones, C. Macroeconomics. 2 Ed W W Norton & Co Inc, , 2011.
Krugman, P & Taylor, L. Contractionary Effects of Devaluations. Working papers 191,
Massachusetts Institute of Technology (MIT), Department of Economics, 1976
Lavoie, M. “A Post Keynesian view of parity theorems”, Journal of Post Keynesian
Economics, fall, 2000
Lavoie, M. "The reflux mechanism and the open economy" in Rochon, L. & Vernengo,
M. "Credit, Interest Rates and Open Economy: Essays on Horizontalism", Edward Elgar,
2001
Lavoie, M. “Interest Parity, risk premia, and Post Keynesian analysis”, Journal of Post
Keynesian Economics , Winter, 2002-2003
Lavoie, M. Endogenous money: accomodationist. Handbook of Alternative Monetary
Economics. Edward Elgar Publishing, 2005
Lavoie, M. ; Daigle, G. A behavioural finance model of exchange rate expectations
within a stock-flow consistent framework”, Metroeconomica, 2011
Mankiw, N. Macroeconomics. 7 ed. , Worth Publishers: NY 2010
Medeiros, C. ; Serrano, F. Capital flows to emerging markets: a critical view based on
the brazilian experience. In: Matias Vernengo. (Org.). Monetary Integration And
Dollarization: No panacea. : Edward Elgar, 2006.
Pivetti, M. An essay on money and distribution. London: MacMillan, 1991.
Pivetti, M. "Monetary endogeneity and non-neutrality in a sraffian perspective" in in
Rochon, L. & Vernengo, M. Credit, Interest Rates and Open Economy: Essays on
Horizontalism, Edward Elgar, 2001
Romer, D. Keynesian macroeconomics without the LM curve, Journal of
EconomicPerspectives, 14 (2), 149-169, 2000.
Romer, D. Short-Run Fluctuations, in:[http://elsa.berkeley.edu/~dromer], Jan. 2006
Santos , G. Uma releitura da teoria das finanças funcionais, tese de doutorado, IE-UFRJ,
2005
Serrano, F. Dollarisation in Latin America. em Pettifor, A. (org) Real world economic
outlook.Palgrave, United Kingdom 2003.
Simonsen, M. ; Cysne, R Macroeconomia, 1 ed. Rio de Janeiro: Ao Livro Técnico, 1989
Smithin, J. “Interest parity, purchasing power parity, “risk premia”, and Post Keynesian
economic analysis”, Journal of Post Keynesian Economics, winter, 2002-2003
Taylor, J. B. A Core of Practical Macroeconomics. American Economic Review, 87 (2),
pp. 233 –235, May 1997.
Taylor, J. B. Teaching modern macroeconomics at the principles level, American
Economic Review, 90 (2), pp. 90–4, 2000
Tcherneva,P. Modern Monetary Theory and Mr. Paul Krugman: a way forward,
disponível em http://neweconomicperspectives.blogspot.com/2011/03/modernmonetary-theory-and-mr-paul.html, 2011
Tsiang, S. C. The Dynamics of International Capital Flows and Internal and External
Balance The Quarterly Journal of Economics Vol. 89, No. 2, May, 1975
Weeks, J. The Effectiveness of Monetary Policy Reconsidered, PERI Working paper 202,
2009
22
Young, W.; Darity, W. "IS-LM-BP: An Inquest"History of Political Economy 36 (Suppl 1):
127–164, 2004.
23